In an indoor communication scenario, a distance between indoor base stations is usually very small (within 30 m). If each base station in an area is configured as an independent cell, interference between the cells is severe. Although a single frequency network configuration may cancel interference between independent cells, spectrum utilization is low.
Multi-user beamforming (MU-BF) is a means for improving spectrum utilization, but can hardly be applied due to limitations of existing products. This is because an existing single base station generally has only two transmit antennas due to a limited volume, and has a limited degree of freedom in space, and it is difficult to find an appropriate multi-user weight. It is found through research that, using multiple base stations jointly for transmission can increase a quantity of transmit antennas, increase the degree of freedom in space, and therefore find multiple orthogonal weights, and can further implement multi-user spatial multiplexing and improve spectrum utilization.
Each radio frequency channel in a remote radio unit (RRU) configured to receive and transmit signals in a base station is made up of independent hardware. This causes responses introduced on uplink and downlink channels in each radio frequency channel to be different, and affects performance of downlink beamforming. Therefore, the uplink and downlink channels in each radio frequency channel need to be corrected by means of channel correction, so that ratios of channel response estimated values of uplink channels to channel response estimated values of downlink channels in all radio frequency channels in the RRU are the same, and that accuracy of downlink beamforming is ensured as much as possible.
Through research and practice, the inventor of the present application finds that although there are some methods for channel correction between multiple RRUs currently, generally two steps are required for implementing channel correction between the multiple RRUs in the existing methods. First, channel self-correction in each RRU needs to be implemented by using a dedicated correction reference channel in each RRU, and then channel correction between the RRUs is performed. For an RRU that has no correction reference channel (namely, an RRU that has no channel self-correction function), the existing channel correction mechanism cannot implement channel correction between multiple RRUs.